The client is a US health center. Since the COVID-19 outbreak, the center has been bursting past capacity. To combat the virus, they decided to go for biometric face recognition time attendance software.
The COVID-19 pandemic has amplified deficiencies of the client’s health center. To combat the virus, they have enhanced precautions and provided the front-line care team with personal protective equipment (PPE). Cleaning and disinfecting touched surfaces lowered the chance of the virus spreading but didn’t solve the problem. In this case, biometric touchless authentication could become a solution. Changing the way healthcare workers clock in and out could decrease the spread of the virus. They needed a solution that would allow masked face recognition. To get a consultation on facial recognition time clock software, they contacted Javed Khattak, an AI and facial recognition service provider.
The client emphasized that they needed a real-time smart attendance system using face recognition techniques. The key focus should be on masked face recognition because the healthcare team at the centre is required to wear masks.
During the project planning phase, we gathered all the client’s requirements, estimated time and risks and came up with a set of plans.
To start, we researched masked face datasets and 80+ of open-source solutions related to face mask detection.
Our team of engineers defined the workflow of the future face recognition time attendance software:
Then, we collected 800+ health center employees’ images, sorted and labelled them with names. The camera at the center’s entrance captured face data and sent it to the server for image processing – detection, encoding, and recognition.
Mask Face Detection
Our team researched the latest studies and decided to use AI and ML for face mask detection and recognition.
We trained our solution to recognize face attributes like:
Detecting and recognizing these facial features, the system can verify the employee’s identity with a mask on.
How the system works:
Within the beta-testing period, we were enhancing face recognition accuracy by experimenting with different types of masks, wrong mask-wearing, bad lighting, and multiple people in the frame.
As a result, we’ve achieved a 91% accuracy in masked face detection.
Our solution is a custom face recognition attendance system with a mask detection feature. It allows for touchless authentication and serves the health centre amid the coronavirus crisis. Thanks to that, employees wearing masks can easily travel in and out of facilities. The solution also provides clock in/out time entry.
Benefits of a Real-Time Face Recognition Time Attendance System: